NCI to enlist chatbots for genetic testing counseling
- By Anoushka Deshmukh
- Jul 24, 2019
When people are diagnosed with a hereditary disease, doctors urge their family members to get tested for pathogenic variants that could identify cancers and other diseases while they are easier and less expensive to treat. This cascade testing first identifies close relatives at risk and then moves to more distant relatives as disease carriers are found.
The current method for cascade testing is vague and outdated, according to the National Cancer Institute. It relies on the individual with the pathogenic variant to notify all at-risk family members and encourage them to get tested. This method has proved ineffective, with only 20% to 52% of at-risk relatives being tested.
The NCI Genetics Branch wants to optimize the cascade testing method with an artificial intelligence-enabled chatbot that converses with family members about testing. It would use sophisticated natural language processing systems to automate text-based conversations that not only educate the user and answer questions but also collect important information prior to a testing appointment. A transcript of each chatbot encounter with a patient would be stored in a secure database and transferred to the NCI Genetics Branch.
NCI is seeking a chatbot that includes the following features:
- Editing that allows NCI Genetics Branch staff to update the conversation script as evidence changes and new input from users is incorporated.
- A basic template script that can be modified and easily autopopulated for the mode of inheritance or specific gene.
- Notifications from the chatbot regarding invitations to patients and reminders to respond.
- Compliance with all Health Insurance Portability and Accountability Act rules and regulations for every patient interaction.
- Functions that allow patients to easily share testing results with relatives or partners and that track if results were shared and how often.
- A smart FAQ that uses natural language understanding and processing to provide appropriate responses to common questions.
The NCI Genetics Branch will compare the chatbot's performance to traditional methods on first-, second- and third-degree relatives at risk.
Responses are due by July 29. Read the full solicitation here.
Anoushka Deshmukh is an intern with Public Sector 360, writing for GCN, FCW and Defense Systems.